When good signals go bad The 2nd Russian banking failure via Mark - - PowerPoint PPT Presentation

When good signals go bad

The 2nd Russian banking failure

L´ ıvia Mark´

• czy

Andrew Spicer

Livia.Markoczy@ucr.edu Andrew.Spicer@ucr.edu

The A. Gary Anderson Graduate School of Management University of California, Riverside

When good signals go bad – p.1/24

Main argument

Some practices that work well in the West may go awry when they applied in a country with weak or underdeveloped institutions. We argue that signals that are relatively accurate reflections of quality in the West may not only not work in other environments but may be the ones that are most likely to be transmitted by counterfeits or mimics.

When good signals go bad – p.2/24

Limits on transferring practices

Constraints:

Cultural Hofstede (1980); Welsh et al. (1993) Legal and regulatory Kostova (1999); Kostova and

Roth (2002); Kostova and Zaheer (1999); Khanna and Palepu (1997)

When good signals go bad – p.3/24

But

Sometimes locals are eager to adopt a practice and there are no cultural or legal barriers to implementing it, but things still go wrong.

When good signals go bad – p.4/24

This can happen when. . .

This can happen when a practice depends on features of an institutional environment which is not present in the new location. We argue based on data from 199 Russian banks that this is what happened in the Russian banking system at the end of nineteen-nineties.

When good signals go bad – p.5/24

Quality and Signals

Theoretical framework: signaling theory Investors tend to look for high quality partners, products and services for their money. An economic partner is high quality if it is accountable, have a long term earning potential and trustworthy (competence and integrity)

When good signals go bad – p.6/24

Quality and Signals, cont.

Quality is often not directly observable, but needs to be derived from indicators (signals). By signal we mean a piece of information that is conveyed by an entity concerning its quality that is not directly observable (Gambetta 2004).

When good signals go bad – p.7/24

Some notation

An Orginator, O, sends a Signal, S, which is received by a Recipient R. There is a Cost, C(O, S), to O for sending S. And there is an Interpertation cost, I(R, S) to R for interperting S.

When good signals go bad – p.8/24

Quality signals of quality

A useful signal of quality is one which

• 1. Is cheap to interpert (I(R, S) is small).
• 2. Is much cheaper for an honest sender to

send than for a dishonest one to send (C(Oh, S) ≪ C(Od, S)).

• 3. Sending and interperting signals does not

have to be deliberate or conscious.

When good signals go bad – p.9/24

Signals of Quality

Amazon.com took out expensive print advertising in the early days of on-line retail Michael Spence has an advanced degree from a prestigious university (and a Nobel Prize, too) That used car comes with a substantial warrentee This peacock has a very healthly looking tail That bower bird has decorated his bower with lots of blue things

When good signals go bad – p.10/24

The problem of mimicry

False signalers may fool “signal receivers" by mimicking signals of quality (Akerlof 1970, Spence 1973) Mimicry here is not simply about copying best

• r legitimate practices, but is rather about

fooling transaction partners about one’s quality Problem for honest signalers: how to distinguish themselves from the mimics

When good signals go bad – p.11/24

The wasp stings; others mimic

Wasp Moth Beetle Fly

Courtesy Peregrine Productions http://www.bombus.freeserve.co.uk/mimicry.htm

When good signals go bad – p.12/24

Two equilibria

Distinguishing equilibrium when honest signalers

have signals that allows them to distinguish themselves from the mimics

Pooling equilibrium when no usable signals exist

that can distinguish honest signalers from the mimics

When good signals go bad – p.13/24

The main points

Salient parts of the signaling system by which westerners judged the viability of banks were imported into Russia. This system failed miserably because the practice tacitly relied

• n institutional conditions not present in

Russia. When importing a successful practice it is important to have a deep understanding of why the practice works, and what its success depends on.

When good signals go bad – p.14/24

Signals and institutions

In the West well developed legal and regulatory conditions (institutions) increase the cost of mimicry (e.g., accounting standards) and aid the honest signalers in developing signals that distinguishes them from the mimics (e.g., auditing, accreditation). These institutions are not always available elsewhere, and thus signals that are relatively accurate reflection of quality in the West may not be so in other countries.

When good signals go bad – p.15/24

Western signals of quality

Costly advertising (signal long term

• rientation)

Ratings by a credible third party

When good signals go bad – p.16/24

Mimicry in Russia

Costly advertising is possible if most money is spent on advertising without spending money

• n developing capabilities. Thus, banks

intending to “take the money and run” can spend more on advertising than honest banks. Without clear accounting standards and enforcing agencies third parties may need to base their information on unreliable sources. The banks can lie about their books. Those ready to lie will have better looking books.

When good signals go bad – p.17/24

Western investors (lenders)

We hypothesize that Western investors rely on Advertising European Bank for Reconstruction & Development (EBRD) rating State scrutiny (tailored regulation) Rating Information Center (RIC) to make investment decisions.

When good signals go bad – p.18/24

Local investors

We hypothesize that local investors rely on Western investors EBRD rating State scrutiny (tailored regulation) Rating Information Center (RIC)

When good signals go bad – p.19/24

Banks that lost their licences

We hypothesize that banks that advertised most, had the highest EBRD ratings, participated in stricter state regulation, and rated highest by RIC were most likely to lose their licences, even after controlling for their founding year, size, and ownership profile.

When good signals go bad – p.20/24

Western Lending

Predictor Step 1 Step 2 Step 1 Assest Size .487∗∗∗ .233∗∗∗ Founding year −.143∗ −.040 Ownership .061 .049 Step 2 Advertising .343∗∗∗ RIC Rating .052 Regulation .229∗∗∗ EBRD .156∗∗ ∆R2 .260 ∆F 27.985∗∗∗ Overall R2 .269 .556 F 27.326∗∗∗ 34.185∗∗∗ df (regress/residual) 3/195 7/191

N = 199; p∗ < .05; p∗∗ < .01; p∗∗∗ < .001

When good signals go bad – p.21/24

Local individual investment

Predictor Step 1 Step 2 Step 1 Assest Size .226∗∗∗ .251∗∗∗ Founding year −.251∗∗∗ −.228∗ Ownership .003 .011 Step 2 Advertising .116 RIC Rating .098 Regulation .245∗∗ Western Loans .157∗ EBRD .183∗ ∆R2 .078 ∆F 3.806∗∗ Overall R2 .143 .221 F 10.826∗∗∗ 6.730∗∗∗ df (regress/residual) 3/1995 7/190

N = 199; p∗ < .05; p∗∗ < .01; p∗∗∗ < .001

When good signals go bad – p.22/24

Logistic Regressing: Losing a license

Predictor Step 1 Step 2 Step 1 Assest Size .209∗∗ .073 Founding year −.200∗∗ −.152∗ Ownership .060 .070 Step 2 Advertising .146∗ RIC Rating .100 Regulation .139† EBRD .150∗ ∆χ2 13.101∗∗ Overall χ2 18.544∗∗∗ 31.645∗∗∗ df 3 7

N = 199; p† < .10; p∗ < .05; p∗∗ < .01; p∗∗∗ < .001

When good signals go bad – p.23/24

Moral

You can not just transport some salient features

• f successful practices without deep

understanding of why these practices work in the first place.

When good signals go bad – p.24/24